Parkin Insufficiency Accentuates High-Fat Diet-Induced Cardiac Remodeling and Contractile Dysfunction Through VDAC1-Mediated Mitochondrial Ca2+ Overload.

Mitochondrial Ca2+ overload contributes to obesity cardiomyopathy, yet mechanisms that directly regulate it remain elusive. The authors investigated the role of Parkin on obesity-induced cardiac remodeling and dysfunction in human hearts and a mouse model of 24-week high-fat diet (HFD) feeding. Parkin knockout aggravated HFD-induced cardiac remodeling and dysfunction, mitochondrial Ca2+ overload, and apoptosis without affecting global metabolism, blood pressure, and aortic stiffness. Parkin deficiency unmasked HFD-induced decline in voltage-dependent anion channel (VDAC) type 1 degradation through the ubiquitin-proteasome system but not other VDAC isoforms or mitochondrial Ca2+ uniporter complex. These data suggest that Parkin-mediated proteolysis of VDAC type 1 is a promising therapeutic target for obesity cardiomyopathy.

Obesity is reaching epidemic proportions globally, severely affecting quality of life and health care, triggering dyslipidemia, insulin resistance, diabetes mellitus, hypertension, and sleep apnea.1, 2, 3 Uncorrected obesity prompts unfavorable changes in cardiac structure and function, including left ventricular (LV) hypertrophy, interstitial fibrosis, hemodynamic alterations, and diastolic dysfunction, a pathologic condition termed obesity cardiomyopathy. The pathogenesis of obesity cardiomyopathy remains elusive, making it challenging to engage effective measures besides weight reduction to halt progressive cardiac remodeling and dysfunction. Recent evidence has demonstrated an essential role for disturbance in mitochondrial Ca2+ handling, especially Ca2+ overload, in deranged redox balance, energy metabolism, cell survival, and cardiac function in the context of obesity.,, Robust mitochondrial Ca2+ influx is evident in the heart through physical contact sites between outer mitochondrial membrane and sarcoplasmic reticulum, namely, mitochondrial-associated membranes where voltage-dependent anion channels (VDACs) connect with sarcoplasmic reticulum Ca2+ channels to ease rapid Ca2+ transport from sarcoplasmic reticulum into mitochondria. VDAC1-mediated mitochondrial Ca2+ import may also present at mitochondria-lysosome contact sites. Ca2+ enters the mitochondrial matrix via mitochondrial Ca2+ uniporter (MCU) located within the inner mitochondrial membrane. Both VDAC1 and the MCU complex are deemed vital for mitochondrial Ca2+ overload and oxidative stress–induced apoptosis.

Recent evidence has depicted a major role for mitophagy (ie, engulfment of long-lived or damaged mitochondria in the double-membraned autophagosomes) in the maintenance of mitochondrial integrity and function, including mitochondrial Ca2+ import, under pathologic stresses., Parkin, an E3 ubiquitin ligase recruited to depolarized mitochondria to ubiquitinate outer mitochondrial membrane proteins, helps remove impaired mitochondria through the mitophagy machinery and promotes the proteasome-dependent degradation of mitochondrial proteins. Nonetheless, changes of mitophagy in advanced stages of obesity and how Parkin regulates mitochondrial Ca2+ import proteins remain at large. Given the pivotal role of Parkin in mitochondrial integrity, mitochondrial Ca2+ load and cardiomyocyte survival, this study was designed to evaluate the effect of Parkin ablation on global metabolism, cardiac remodeling, and contractile and aortic function in high-fat diet (HFD)–induced obesity and possible mechanisms involved. Results from our study reveal that Parkin deficiency accentuates cardiac remodeling, contractile (systolic and diastolic) dysfunction, mitochondrial impairment, and mitochondria-mediated cell death through mitochondrial Ca2+ overload without affecting global metabolic dysfunction, aortic stiffness, and elevated blood pressure (BP) in diet-induced obesity, denoting a vital role for Parkin in the regulation of mitochondrial Ca2+ influx.

Methods

Bioinformatics analysis

Messenger RNA profiles of human hearts (specifically atrial biopsies) were downloaded from the Gene Expression Omnibus database (GSE159612), including 9 lean (body mass index 22.2 ± 0.6 kg/m2) and 9 obese (body mass index 33.7 ± 1.1 kg/m2) patients (information presented in Supplemental Table S1). Gene Ontology analysis and gene set enrichment analysis were performed in R version 4.1.1 (see the Supplemental Methods for details).

Human samples

LV samples were obtained from unsuccessful cardiac transplants from 8 lean (body mass index 19.6 ± 1.2 kg/m2) and 8 obese (body mass index 31.8 ± 1.2 kg/m2; P < 0.05 vs lean group) donors (Supplemental Table S2). The protocol was approved by the Sun Yat-Sen Memorial Hospital Ethics Committee (SYSEC-KY-KS-2019-019) and was in line with the principles defined in the Declaration of Helsinki.

Experimental animals and HFD feeding

The experimental procedures described here were approved by the Institutional Animal Use and Care Committees of Zhongshan Hospital Fudan University and were in accordance with National Institutes of Health guidelines. In brief, wild-type (WT) and global Parkin knockout (Parkin−/−) mice (B6.129S4-Parktm1Shn/J, strain 006582) on C57BL/6J background were purchased from the Jackson Laboratory (see the Supplemental Methods for details). Two-month-old male WT and Parkin−/− mice were randomly assigned to a low-fat diet (LFD; 10% of calories from fat; D12450H, Research Diets) or an HFD (45% of total calories from fat; D12451, Research Diets) for 24 weeks. Mice were anesthetized using ketamine (80 mg/kg; Pfizer) and xylazine (12 mg/kg; Bayer) before being sacrificed using cervical dislocation.

Metabolic cage

Mice were individually housed and monitored in a semisealed metabolic monitoring system (CLAMS, Columbus Instruments) at 22°C with free access to food and water for 24 hours (7 AM to 7 AM the next day). Parameters of O2 consumption, CO2 production, respiratory exchange ratio (CO2 production/O2 consumption), heat production ([(3.815 + 1.232 × respiratory exchange ratio) × O2 consumption] × 1,000), and voluntary activity were determined for a 24-hour period.

Plasma file

Blood glucose and plasma insulin levels were measured using a glucometer (Bayer) and enzyme-linked immunosorbent assay commercial kits (R & D Systems). Homeostasis model assessment–estimated insulin resistance index was calculated as: [fasting insulin (mU/L) × fasting glucose (mmol/L)]/22.5.

Histologic examination:

Following anesthesia, mice were sacrificed and hearts were excised and immediately placed in 10% neutral-buffered formalin at room temperature for 24 hours. Specimens were embedded in paraffin, cut into 5-μm sections, and stained with fluorescein isothiocyanate–conjugated wheat germ agglutinin. Oil red O and Masson’s trichrome staining was used to detect lipid and fibrosis, respectively. Images were captured using an Olympus BX-51 microscope (Olympus America) and analyzed using the Image J Fiji version 2.3.0 (National Institutes of Health).,

Cardiac and vascular function assessment

Mice were anesthetized prior to evaluation of cardiac and aortic function using 2-dimensionally guided M-mode and color Doppler echocardiography (Vevo 2100, FUJIFILM VisualSonics) equipped with a 22- to 55-MHz linear transducer (MS550D, FUJIFILM VisualSonics). Ejection fraction and heart rate were derived using Vevo 2100 echocardiography., LV diastolic function was assessed using the ratio of early to late ventricular filling velocities (E/A) and deceleration time. Systolic and diastolic BP were measured using a CODA semiautomated noninvasive device (Kent Scientific). The distance between the aortic arch and the bifurcation of left common carotid artery divided by pressure-wave transit time was used to generate pulse-wave velocity (see Supplemental Methods for details).

Isolation and treatment of mouse cardiomyocytes

Adult murine cardiomyocytes (AMCMs) were collected from diet-fed male WT or Parkin−/− mice to examine contractile function and Ca2+ levels. In addition, a cohort of AMCMs were treated with the proteasomal inhibitor MG132 and autophagy inhibitors (3-methyladenine and bafilomycin A1) to discern degradation modalities. AMCMs were also treated with palmitic acid (PA) in the absence or presence of the VDAC1 inhibitor or the mitochondrial permeability transition pore (mPTP)–opening inducer. Neonatal mouse cardiomyocytes were isolated from 1-day-old WT male mice using a series of enzymatic digestions (Worthington Biochemical) and were transduced with adenoviruses or small interfering RNA for mitophagy detection (see Supplemental Methods for details).

Cell shortening and relengthening

Mechanical properties of AMCMs were assessed using an IonOptix soft-edge system. Cardiomyocytes were field-stimulated at 0.5 Hz. Cell shortening and relengthening were assessed including peak shortening, time to peak shortening, time to 90% relengthening (TR90), and maximal velocities of shortening and relengthening (±dL/dt).,

Measurement of intracellular and mitochondrial Ca2+

AMCMs were loaded with Fura-2/AM (0.5 μM; F1201, Thermo Fisher Scientific) for 15 minutes, and fluorescence intensity was recorded with a dual-excitation fluorescence photomultiplier system (IonOptix). Qualitative change in Fura-2 fluorescence intensity was inferred from the Fura-2 fluorescence intensity ratio at the 2 wavelengths (360 and 380 nm). Fluorescence decay time (single exponential) was calculated as an indicator of intracellular Ca2+ clearance. Mitochondrial Ca2+ transients were monitored using Rhod-2 AM (1 μM, 1 hour at 37°C; R1245MP, Thermo Fisher Scientific), followed by de-esterification for another hour in Rhod-2-free medium. Cells were exposed to light emitted by a 75-W lamp while being stimulated to contract at a frequency of 0.5 Hz. Mitochondrial Ca2+ transients were presented as background-subtracted normalized fluorescence. In addition, a cohort of AMCMs were imaged through a confocal microscope (TCS SP8, Leica) with fluorescence excitation and emission maxima at 552 and 581 nm. The fluorescence intensity of Rhod-2 was analyzed using Image J Fiji version 2.3.0.

Transmission electron microscopy

Small cubic pieces ≤1 mm3 were dissected from left ventricles and were fixed with 2.5% glutaraldehyde in 0.1 M sodium phosphate (pH 7.4) overnight at 4°C. Following postfixation in 1% OsO4, samples were dehydrated through graded alcohols and were embedded in Epon Araldite. Ultrathin sections (50 nm) were cut using an ultramicrotome (UltraCut E, Leica) and stained with uranyl acetate and lead citrate. Images were captured using the FEI Tecnai G2 Spirit transmission electron microscope.

Reactive oxygen species detection

Superoxide anion (O2−) was detected in heart sections (30 μm thick) using dihydroethidium staining (2 μM; Invitrogen) for 30 minutes in a humidified chamber at 37°C. Images were obtained using a computer-assisted microscope. For the detection of reactive oxygen species (ROS), isolated AMCMs were incubated with MitoSOX red mitochondrial superoxide indicator (5 μM; M36008, Thermo Fisher Scientific) for 30 minutes and were imaged using confocal microscopy.

Detection of mitophagy, mitochondrial function, and apoptosis

Neonatal mouse cardiomyocytes were transfected with adenoviruses encoding mt-Keima (multiplicity of infection = 50; Hanbio Biotechnology). Levels of mitophagy were quantified as the ratio of fluorescence signal area at 561-nm excitation to signal derived from 458-nm excitation with a 570- to 695-nm emission range using confocal microscopy. Details of additional methods and materials, including mitochondrial aconitase activity, NAD+ levels, and western blotting are described in the Supplemental Methods.

Statistical analysis

Data are expressed as mean ± SEM. Statistical significance (P < 0.05) was determined using 1-way or 2-way analysis of variance followed by the Tukey post hoc test for multiple pairwise comparisons. Statistical analyses were performed using GraphPad Prism version 9.0.

Results

Obesity evoked a decline in mitochondrial turnover and quality in human heart

To discern biological processes in obese hearts, publicly available RNA sequencing result (GSE159612) from the Gene Expression Omnibus database containing myocardial tissues from subjects with different body mass index was analyzed. Gene Ontology analysis showed that the differentially expressed genes between heart samples from obese patients and normal subjects were overtly enriched in biological processes associated with intracellular Ca2+ homeostasis, including “response to calcium ion” and “calcium-ion regulated exocytosis” (Figure 1A). Gene set enrichment analysis was used to determine whether a series of prior defined sets of genes (eg, those from a specific Gene Ontology term or Kyoto Encyclopedia of Genes and Genomes pathway) was concordantly different between lean and obese individuals. Gene set enrichment analysis revealed that genes encoding Ca2+ import into the mitochondria were consistently more abundant in obese than lean individuals, whereas other biological processes associated with Ca2+ regulation, such as “Ca2+ ion import into cytosol,” were down-regulated in obese compared with normal-weight human hearts (Figures 1B to 1D).

Figure 1

Enrichment Analysis Result of Transcriptome and Mitochondrial Protein Profiles in the Human Heart

(A) Bubble plot showing top enriched Gene Ontology (GO) terms of biological process. Gene set enrichment analysis (GSEA) result of (B) “Ca2+ import into mitochondria” and (C) “Ca2+ ion import into cytosol.” Distribution of genes in the ranked gene list is shown as “hits.” The enrichment score (green line) represents the running enrichment score at any point in the gene set. The bottom portion exhibits the value of ranking metric that measures a gene’s correlation with a phenotype. P values and the normalized enrichment score (NES) are presented in tabular format. (D) Ridge plot showing GSEA result of enriched biological processes (BPs) related to Ca2+ (P < 0.05, |NES| > 1). (E) Mitochondrial or mitophagy proteins (Parkin, LC3BII-I, p62, PGC1α, UCP2, and VDAC1) in lean and obese human hearts. GAPDH, α-tubulin, or COX IV was used as the loading control. Mean ± SEM, n = 6 to 8 samples/group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 between indicated groups.

Next, levels of proteins associated with mitochondrial Ca2+ and mitochondrial quality control were evaluated in human heart samples. Consistent with the gene set enrichment analysis results, levels of VDAC1 were up-regulated along with down-regulated LC3BII/LC3BI ratio, Parkin, peroxisome proliferator–activated receptor γ coactivator 1α (a transcriptional coactivator of mitochondrial biogenesis), and uncoupling protein 2 (a predominant isoform of mitochondrial uncoupling protein), as well as increased accumulation of p62 in obese human hearts (Figure 1E). These data provide evidence for dampened mitochondrial turnover (autophagy and biogenesis) and buildup of functionally compromised mitochondria in obese hearts.

Parkin deficiency did not affect global energy metabolism in response to HFD intake

Parkin−/− and WT mice were fed an HFD for 24 weeks. The HFD evoked metabolic derangement, shown as decreased O2 consumption, CO2 production, respiratory exchange ratio, and heat production, without affecting physical activity, in a comparable manner in both WT and Parkin−/− mice (Figures 2A to 2P). The lower energy expenditure in HFD-fed mice was not related to decreased physical activity (Figures 2O and 2P). These findings suggested that HFD intake switched energy source with permission from carbohydrates to fats in WT and Parkin−/− mice. Thus, Parkin deficiency does not affect HFD intake–induced global metabolic disorder, ruling out global impact of Parkin on energy metabolism following HFD intake.

Figure 2

Metabolic Properties (During a 24-Hour Cycle) in WT and Parkin−/− Mice Fed an LFD or HFD for 24 Weeks

(A) O2 consumption (Vo2; 24 hours). (B) Pooled Vo2 (light). (C) Pooled Vo2 (dark). (D) Pooled Vo2 (combined). (E) CO2 production (Vco2; 24 hours). (F) Pooled Vco2 (light). (G) Pooled Vco2 (dark). (H) Pooled Vco2 (combined). (I) Respiratory exchange ratio (RER; 24 hours). (J) Pooled RER (light). (K) Pooled RER (dark). (L) Pooled RER (combined). (M) Heat production (24 hours). (N) Pooled heat production. (O) Total physical activity (24 hours). (P) Pooled total physical activity. Mean ± SEM, n = 7 or 8 mice/group. ∗P < 0.05 vs WT-LFD group (A, E, I, M, O); ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 between indicated groups (B to D, F to H, J to L, N, P). HFD = high-fat diet; LFD = low-fat diet; WT = wild-type.

Biometrics, insulin resistance, and cardiac geometry of mice fed an LFD or HFD

Consistent with metabolic cage data, body weight was comparably increased in WT and Parkin−/− mice receiving the HFD (Figure 3A). HFD consumption equally increased plasma levels of insulin and triglycerides without affecting fasting blood glucose levels in WT and Parkin−/− mice (Figures 3B to 3D). Homeostasis model assessment–estimated insulin resistance index exhibited comparably impaired insulin sensitivity in WT and Parkin−/− mice receiving the HFD (Figure 3E). Interestingly, HFD-induced increases in gross heart weight were more pronounced in Parkin−/− mice compared with WT mice (Figure 3F), indicating a potential role of Parkin in the maintenance of cardiac geometry in obesity. Although heart weight normalized to tibial length failed to reach statistical significance in HFD groups (Figure 3G), Parkin ablation accentuated HFD intake–provoked rises in cardiomyocyte cross-sectional area, intracardial accumulation of lipids, and interstitial fibrosis (Figures 3H to 3M). These data indicate that Parkin deficiency accentuates HFD intake–induced cardiac remodeling and steatosis without affecting overall insulin sensitivity.

Figure 3

HFD Intake–Induced Body Weight Gain, Glucose Intolerance, Plasma Profile, and Cardiac Remodeling in WT and Parkin−/− Mice

(A) Body weight gain over the 24-week feeding period. (B) Fasting blood glucose. (C) Serum triglycerides. (D) Plasma insulin. (E) Homeostasis model assessment–estimated insulin resistance (HOMA-IR) index. (F) Heart weight. (G) Heart weight normalized to tibial length. (H) Representative images of lectin. (I) Pooled cardiomyocyte cross-sectional area. (J) Representative images of Oil red O. (K) Pooled quantitative analysis of lipid content. (L) Representative images of Masson trichrome staining. (M) Pooled quantitative analysis of interstitial fibrotic area (as a percentage of entire cardiac region). Mean ± SEM, n = 5 to 7 mice/group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 between indicated groups. Abbreviations as in Figure 2.

Parkin ablation worsened HFD-induced cardiac dysfunction without affecting vascular function

Changes in cardiac systolic and diastolic function, cardiac geometry, and aortic stiffness were monitored using echocardiography. Although Parkin ablation alone did not affect echocardiographic parameters, it augmented HFD intake–induced increases in LV end-systolic diameter and unmasked HFD-induced increase in LV end-diastolic diameter, in line with the exacerbated decline in fractional shortening and ejection fraction as well as increases in LV mass with HFD intake, without any notable changes in LV wall thickness, septal thickness, and heart rate (Figures 4A to 4I). Parkin removal accentuated HFD-induced LV diastolic defect, as evidenced by a more pronounced decline in E/A ratio (Figure 4J). However, deceleration time remained similar in all groups (Figure 4K). In addition, HFD increased systolic but not diastolic BP (Figures 4L and 4M) and promoted arterial stiffness (Figures 4N and 4O) (as manifested by higher pulse-wave velocity values), the effects of which were unaffected by Parkin ablation. Parkin knockout itself did not affect BP and aortic stiffness (Figures 4L to 4O). These findings denote that Parkin ablation augments HFD intake–induced cardiac dysfunction and remodeling without affecting BP and aortic stiffness.

Figure 4

HFD Intake–Induced Changes in Echocardiographic Indexes in WT and Parkin−/− Mice

(A) Representative echocardiographic images from 4 mouse groups. (B) Left ventricular (LV) wall thickness. (C) Septal thickness. (D) LV end-systolic diameter (LVESD). (E) LV end-diastolic diameter (LVEDD). (F) Fractional shortening. (G) Ejection fraction. (H) LV mass. (I) Heart rate. (J) E/A ratio. (K) Deceleration time (DT). (L) Diastolic blood pressure (BP). (M) Systolic BP. (N) Representative images of pulse-wave velocity (PWV). (O) PWV. Mean ± SEM, n = 7 to 9 mice/group. ∗P < 0.05, ∗∗P < 0.01, or ∗∗∗P < 0.001 between indicated groups. Abbreviations as in Figure 2.

Parkin ablation aggravated cardiomyocyte contractile and intracellular Ca2+ derangement in the face of HFD intake

Neither the HFD nor Parkin ablation exhibited any notable responses on cardiomyocyte cell length (Figure 5A). In comparison with LFD intake, HFD intake induced decreased peak shortening, maximal velocity of shortening and relengthening (±dL/dt), and prolonged relengthening duration (TR90) without affecting shortening duration (time to peak shortening), which were more pronounced in Parkin−/− mice compared with WT mice (Figures 5B to 5F). To decipher potential mechanisms underlying Parkin ablation–elicited deterioration of HFD-induced cardiomyocyte dysfunction, intracellular Ca2+ handling was evaluated using the Ca2+-sensitive fluorescence dye Fura-2. Parkin ablation overtly accentuated the HFD-provoked rise in basal intracellular Ca2+ level, decrease of the electrically stimulated increase in intracellular Ca2+ (change in Fura-2 fluorescence intensity), and prolongation of intracellular Ca2+ clearance without any notable effect itself (Figures 5G to 5I).

Figure 5

Effect of HFD Intake on Cardiomyocyte Contractile and Intracellular Ca2+ Properties in WT and Parkin−/− Mice

(A) Resting cell length. (B) Peak shortening (percentage of resting cell length). (C) Maximal velocity of shortening (+dL/dt). (D) Maximal velocity of relengthening (−dL/dt). (E) Time to peak shortening (TPS). (F) Time to 90% relengthening (TR90). (G) Baseline Fura-2 fluorescence intensity (FFI). (H) Increase in FFI (ΔFFI) in response to electric stimuli. (I) Intracellular Ca2+ decay rate. Mean ± SEM, n = 56 (A to F) or 40 (G to I) cells from 4 mice per group. ∗P < 0.05, ∗∗P < 0.01, or ∗∗∗P < 0.001 between indicated groups. Abbreviations as in Figure 2.

HFD intake–induced mitochondrial injury and ROS accumulation were accentuated in Parkin−/− mice

Given the role of Parkin-mediated mitophagy in mitochondrial quality control,, mitochondrial morphology and function were examined. Parkin−/− mice displayed greater cytoarchitectural aberrations as a consequence of HFD intake compared with WT mice, such as mitochondrial swelling, fragmentation of cristae, and distortion of sarcomeres and myocardial filaments on transmission electron microscopic analysis, with little ultrastructural change from Parkin ablation itself (Figure 6A). Parkin−/− mice exhibited higher mitochondrial density, depicting accumulation of damaged mitochondria and reduced circulatory index (mitochondrial major axis length/minor axis length), with little change in area per mitochondria in response to HFD intake (Figures 6D to 6F). Likewise, HFD intake promoted mitochondrial ROS production with a more pronounced response in Parkin−/− mice (Figures 6B, 6C, 6G, and 6H). Furthermore, our data demonstrated more pronounced increases in mitochondrial Ca2+ in hearts from Parkin−/− mice following HFD intake compared with WT mice (Figure 6I). Sustained mitochondrial Ca2+ overload triggers mPTP opening, resulting in rapid dissipation of proton gradient across the inner mitochondrial membrane, which drives energetic deficits and cell death.24, 25, 26 Our data revealed that Parkin deletion accentuated HFD-induced mPTP opening, as evidenced by NAD+ levels (Figure 6J), suggesting a possible role for Parkin-mediated mitochondrial Ca2+ overload and mPTP opening in HFD-induced cardiac injury. In line with the loss of mitochondrial aconitase activity (Figure 6K), Parkin deletion accentuated HFD intake–induced changes in PGC1α, UCP2, and TOM20 (Figures 6L to 6O). These data reveal that Parkin deficiency impairs mitochondrial integrity and increased oxidative stress, leading to disruption of mitochondrial function upon HFD consumption.

Figure 6

Effect of HFD Intake on Myocardial Ultrastructure, Mitochondrial O2− Production, Mitochondrial Respiration, and Mitochondrial Integrity in WT and Parkin−/− Mice

(A) Representative transmission electron microscopic (TEM) images of mitochondrial ultrastructure. (B) Representative dihydroethidium (DHE) fluorescence images. (C) Representative MitoSOX fluorescence images. (D) Mitochondrial area (percentage total area). (E) Mitochondrial area (per mitochondrion). (F) Mitochondrial circularity index (major axis length/minor axis length). (G, H) Pooled data of mitochondrial reactive oxygen species levels using DHE staining or MitoSOX staining. (I) Baseline mitochondrial Ca2+ levels using Rhod-2. (J) NAD+ level depicting mitochondrial permeability transition pore opening. (K) Mitochondrial aconitase activity. (L) Representative immunoblots of UCP2, PGC1α, and TOM20. GAPDH or α-tubulin was used as the loading control. Quantified mitochondrial protein levels of (M) UCP2, (N) PGC1α, and (O) TOM20. Mean ± SEM, n = 6 or 7 mice (A to H, J to O) or 15 cells from 4 mice (I) per group. ∗P < 0.05, ∗∗P < 0.01, or ∗∗∗P < 0.001 between indicated groups. Abbreviations as in Figure 2.

HFD-induced changes in autophagy and mitophagy and mitochondria-mediated cell death were augmented by Parkin ablation

Treatment of PA or Parkin knockdown with small interfering RNA significantly suppressed mitophagy, while Parkin overexpression rescued PA-induced loss of mitophagy in AMCMs (Figures 7A and 7B). Parkin deletion augmented HFD-induced autophagy loss with little effect itself (LC3BII/LC3BI ratio and p62) in mice (Figures 7C and 7D). Parkin was slightly suppressed in murine hearts following the 24-week HFD feeding (Figure 7E). HFD evoked overt production of inflammatory cytokines, including IL-6 and TNF-α, the effect of which was further enhanced by Parkin ablation, without any effect from Parkin knockout itself (Figures 7F and 7G).

Figure 7

Effect of HFD Intake on Autophagy, Mitophagy, and Mitochondria-Mediated Cell Death in Hearts From WT and Parkin−/− Mice

(A) Representative images of mt-Keima. (B) Pooled data of mt-Keima fluorescence signal (561 nm/458 nm). Quantified protein levels and representative immunoblots as insets of (C) LC3BI/II, (D) p62, (E) Parkin, (F) IL-6, (G) TNF-α, (H) FasL, (I) caspase-8, (J) Bax, (K) Bcl-2, (L) cytosolic apoptosis-inducing factor (AIF), (M) cytosolic cytochrome c, (N) mitochondrial cytochrome c, (O) caspase-9, and (P) caspase-3. GAPDH, α-tubulin, or COX IV was used as the loading control. Mean ± SEM, n = 10 images (A, B) or 5 to 7 mice (C to P) per group. ∗P < 0.05, ∗∗P < 0.01, or ∗∗∗P < 0.001 between indicated groups. OE = overexpression; PA = palmitic acid; siRNA = small interfering RNA; other abbreviations as in Figure 2.

Regulated cell death involves cell surface death receptors or damaged mitochondria., Our data revealed that death receptor Fas ligand was unaffected by HFD intake, and caspase-8 was comparably up-regulated in the hearts of Parkin−/− and WT mice following HFD intake (Figures 7H and 7I), indicating less likely any involvement of death receptor-mediated apoptosis in Parkin deletion–accentuated anomalies following HFD intake.

Execution of mitochondria-dependent apoptosis involves mitochondrial outer membrane permeabilization, which is tightly regulated by the B cell lymphoma-2 (BCL-2) family of proteins. We noted activation of pro–cell death protein Bax and suppression of prosurvival Bcl-2 following HFD consumption, with a more profound response in Parkin−/− mice (Figures 7J and 7K). Induction of mitochondrial outer membrane permeabilization triggers cell death through release of mitochondrial proapoptotic factors such as cytochrome c and apoptosis-inducing factor. Release of cytochrome c and apoptosis-inducing factor from mitochondria to cytosol was increased following HFD intake, with a more abrupt rise in Parkin−/− mice (Figures 7L to 7N). Parkin ablation augmented activation of caspase-9 and caspase-3 following release of cytochrome c in face of HFD intake (Figures 7O and 7P). Thus, loss of Parkin deteriorates mitochondria-dependent apoptosis following HFD intake.

Parkin regulated HFD-induced mitochondrial Ca2+ overload through proteasome-dependent degradation of VDAC1

Our data revealed that HFD intake led to mitochondrial Ca2+ overload and mPTP opening, culminating in cell death. To this end, levels of VDAC isoforms and MCU complex, which gate import of Ca2+ in outer mitochondrial membrane and inner mitochondrial membrane, respectively,, were examined. Our data demonstrated that only VDAC1 was further up-regulated in hearts from Parkin−/− mice following HFD intake compared with WT mice (Figure 8A), while VDAC2, VDAC3, MCU, MICU1, MICU2, and MCUb were unchanged or equally affected by HFD intake (Figures 8B to 8G). Thus, Parkin deficiency amplified HFD-induced mitochondrial Ca2+ overload, possibly through regulating VDAC1 to stimulate mitochondria-mediated cell death.

Figure 8

Effect Of Parkin On Mitochondrial Ca2+ Transport Protein Levels

(A) Ca2+ channel in the outer mitochondrial membrane VDAC1. (B) VDAC2. (C) VDAC3. (D) Mitochondrial inner membrane Ca2+ uniporter subunit mitochondrial Ca2+ uniporter (MCU). (E) MICU1. (F) MICU2. (G) MCUb. Insets show representative immunoblots depicting protein levels with GAPDH or α-tubulin as the loading control. (H) Scheme depicting protein degradation pathways and inhibitors. (I, J) Adult murine cardiomyocytes were transduced with LacZ or Parkin adenovirus for 48 hours prior to treatment with proteasome inhibitor MG132 (10 μM, 18 hours), autophagy inhibitor 3-MA (5 mM/L, 18 hours), or lysosome inhibitor bafilomycin A1 (50 nM, 4 hours) in the presence or absence of PA (0.5 mM, 8 hours). (I) Representative immunoblots depicting protein levels of VDAC1 and Parkin. α-Tubulin was used as the loading control. (J) Quantified VDAC1 protein level. Mean ± SEM, n = 5 to 7 mice/group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 between indicated groups (A to G) or vs vehicle control (J). ###P < 0.001 vs vehicle-PA-LacZ (J). Abbreviations as in Figures 2 and 7.

Parkin promotes ubiquitination and degradation of mitochondrial proteins. These ubiquitinated proteins were subsequently subjected to proteasome-dependent degradation or removal by autophagy machinery (Figure 8H). VDAC1 is a perceived substrate for Parkin-mediated ubiquitination., We examined the effect of Parkin on VDAC1 in response to PA in mouse cardiomyocytes using Parkin adenoviral transfection. Our data revealed that Parkin effectively rescued the rise in VDAC1 level evoked by PA (0.5 mM for 8 hours) (Figures 8I and 8J, lanes 1 to 3). Next, cardiomyocytes were treated with the proteasome inhibitor MG132, the autophagy inhibitor 3-MA, or the lysosome inhibitor bafilomycin A1 to determine the underlying modalities of Parkin-mediated VDAC1 degradation (Figure 8H). MG132 treatment up-regulated VDAC1 in the absence of PA and abolished Parkin-evoked VDAC1 degradation in the presence of PA, while autophagy inhibition did not affect basal or Parkin-induced changes in VDAC1 levels (Figures 8I and 8J). These results suggest that insufficiency in both Parkin and Parkin-mediated degradation of VDAC1 by the ubiquitin-proteasome system may contribute to HFD-induced mitochondrial Ca2+ overload.

Parkin protected HFD-induced mitochondrial Ca2+ overload and contractile dysfunction

A cohort of adult mouse cardiomyocytes was transfected with Parkin overnight prior to challenge with PA (0.5 mM for 8 hours) in the presence or absence of pharmacologic agents to evaluate cardiomyocyte contractile function and mitochondrial Ca2+ levels. Our data showed that PA significantly elevated mitochondrial Ca2+ levels and impaired cardiomyocyte function, whereas Parkin overexpression or the VDAC1 inhibitor DIDS reversed mitochondrial Ca2+ overload (Figures 9A and 9B) and contractile dysfunction (decreased peak shortening, ±dL/dt, prolonged TR90) induced by PA (Figures 9C to 9H).

Figure 9

Effect of Parkin Transfection and Interventions of VDAC1 or Mitochondrial Permeability Transition Pore Opening on PA-Induced Changes in Ca2+ Homeostasis and Cardiomyocyte Contractile Function

Adult mouse cardiomyocytes were transfected with Parkin using adenovirus (Parkin OE) overnight prior to incubation with PA (0.5 mM) for another 8 hours in the presence or absence of the VDAC1 inhibitor DIDS (200 μM, 1 hour), mitochondrial permeability transition pore opening activator HA14-1 (2 μM, 24 hours), or cisplatin (20 μM, 24 hours). (A) Representative confocal microscopic images of mitochondrial Ca2+ fluorescence in adult murine cardiomyocytes stained with Rhod-2 (red). (B) Quantified Rhod-2 fluorescent intensity. (C) Resting cell length. (D) Peak shortening (percentage of resting cell length). (E) Maximal velocity of shortening (+dL/dt). (F) Maximal velocity of relengthening (−dL/dt). (G) TPS. (H) TR90. Mean ± SEM, n = 10 cells from 4 mice (A, B) or 20 cells from 4 mice (C to H) per group. ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 between indicated groups. Abbreviations as in Figures 2, 5, and 7.

mPTP opening is vital for mitochondria-mediated cell death, where the inner mitochondrial membrane allows free passage of solutes up to 1.5 kDa in size and ultimately triggers cell death. Inhibition of Bcl-2 using HA14-1 sensitizes mPTP opening in high-Ca2+ environments, and cisplatin induces mitochondrial Ca2+ overload and mPTP opening., Although low-dose HA14-1 or cisplatin alone did not exert any effects on mitochondrial Ca2+ load and contractile function, Parkin-evoked benefits on cardiomyocyte mitochondrial Ca2+ and contractile function were obliterated by HA14-1 or cisplatin (Figures 9C to 9H). None of the pharmacologic agents or viral vector affected cardiomyocyte function and mitochondrial Ca2+ levels (Figures 9A to 9H). Taken together, exogenous Parkin expression or VDAC1 inhibition protects cardiomyocytes from PA-induced contractile dysfunction through inhibition of mitochondrial Ca2+ overload and mPTP opening.

Discussion

Our salient findings reveal that Parkin deficiency accentuates HFD-evoked cardiac remodeling and dysfunction, mitochondrial Ca2+ overload, as well as mitochondria-dependent cell death, with no impact on insulin sensitivity, global energy metabolism, aortic stiffness, and BP with either LFD or HFD intake. Furthermore, Parkin-mediated proteasome-dependent degradation of VDAC1 may be responsible for these unfavorable effects. These findings favor a unique role for the Parkin-VDAC1 axis as a therapeutic target in obesity-induced cardiac anomalies, possibly through mitochondrial Ca2+ regulation.

A wealth of research has documented obesity-induced unfavorable myocardial geometric and functional changes, including cardiac hypertrophy, interstitial fibrosis, compromised cardiac contractility, and prolonged diastole, along with higher BP and vascular dysfunction.1, 2, 3 This is consistent with our observations of cardiac hypertrophy (increased LV mass and chamber size), interstitial fibrosis, reduced fractional shortening, ejection fraction, E/A ratio, peak shortening, ±dL/dt, prolonged TR90, elevated BP, and increased pulse-wave velocity following 24-week HFD intake. In general, sustained obesity contributes to cardiomyopathy characterized by diastolic dysfunction. Besides diastolic dysfunction manifested by a decreased E/A ratio, our data also noted that LV fractional shortening and ejection fraction were significantly decreased following 24-week HFD feeding, in line with earlier observations of long-term obesity.

In our hands, a tight link is evident between Ca2+ disturbance and mitochondrial injury in the realm of lipotoxic cardiac dysfunction, with a more pronounced change in Parkin−/− mice. Parkin plays a crucial role as an adaptor in maintaining the integrity of mitochondrial structure and function in the heart. Previous studies have shown that Parkin prevents myocardial ischemia-reperfusion injury and hypoxia-mediated cell death., However, there is an inconsistency with regard to the precise role of mitophagy in HFD intake–induced cardiac injuries. Tong et al found that cardiac autophagic flux in mice fed an HFD peaked at 6 weeks and later declined over time, while elevated mitophagy was observed as early as 3 weeks after the start of the HFD and lasted for 8 weeks. In contrast, we and others have noted declined mitophagy regulators, including Parkin and FUNDC1, in the heart after longer (12 weeks or 20 weeks) HFD feeding. Mitophagy, evaluated by mt-Keima, was also suppressed by PA treatment in isolated cardiomyocytes. Taken together, Parkin and mitophagy seem to be elevated at early stages of HFD-induced heart injury but decline over time.

Given that patients usually manifest cardiac dysfunction after a long period of obesity, declined Parkin may play an essential role at advanced stages of obesity cardiomyopathy. Indeed, our data reveal that Parkin ablation accentuates HFD-induced unfavorable myocardial responses, including cardiac remodeling as well as systolic and diastolic dysfunction. It was reported that Parkin−/− mice were resistant to 6-week HFD feeding–induced body weight gain. However, after long-term HFD feeding, Parkin−/− mice gained weight in our present study. Furthermore, HFD intake–induced insulin resistance, metabolic derangement, elevated BP, and aortic stiffness are not affected by Parkin ablation, ruling out possible contributions from global and vascular effects of Parkin ablation.

A number of scenarios should be considered for Parkin ablation–induced exacerbation of cardiac remodeling and dysfunction under HFD intake. Our data highlights an essential role of Parkin in the regulation of Ca2+ homeostasis, consistent with the Gene Expression Omnibus sequencing data from lean and obese human hearts. The correlation between mitochondrial Ca2+ overload and impaired cardiac function has been extensively examined. Leaky type 2 ryanodine receptors on sarcoplasmic reticulum caused mitochondrial Ca2+ overload and cardiac dysfunction in heart failure. Knockdown of PKD2L1 led to exacerbated mitochondrial Ca2+ overload and cardiac hypertrophy following high salt loading. Overexpression of SERCA2a alleviated cardiac microvascular ischemic injury through suppressing Mfn2-mediated mitochondrial Ca2+ import from endoplasmic reticulum. Decreased Parkin upstream initiator PINK1 in the heart during sepsis led to defective mitochondrial Ca2+ efflux and cardiomyocyte injury. Furthermore, mounting evidence from liver, adipose tissues, and macrophages has illustrated a role for Ca2+ defect in obesity. For example, HFD-induced prolonged elevation of cytosolic Ca2+ was responsible for autophagy defects in the liver. Recently, several attempts have been made to clarify the regulation of HFD-induced Ca2+ overload in the heart. Our laboratory reported that loss of FUNDC1 accentuated HFD-induced mitochondrial Ca2+ overload and cardiac dysfunction through interacting with FBXL2 to govern degradation of the endoplasmic reticulum Ca2+ channel IP3R3. However, the underlying mechanism involved in mitochondrial Ca2+ import, such as the Ca2+ transporter VDAC1 and MCU complex, in the context of obesity remains unclear. Our results also demonstrate a crucial role for mitochondrial ROS production, mitochondrial injury, inflammation, as well as mitochondria-mediated (not death receptor–mediated) cell death in Parkin deficiency–accentuated cardiac injury following HFD intake, denoting a vital role for mitochondrial injury and cell death in Parkin deficiency–induced cardiac anomalies in obesity.

Parkin, as a E3 ligase, directly binds to and ubiquitinates mitochondrial Ca2+ import proteins, including VDAC isoforms and MCU complex.,,, We found that lack of Parkin-mediated degradation may account for the buildup of mitochondrial Ca2+ import protein VDAC1 rather than other isoforms of VDAC or MCU complex. Previous studies have indicated a pivotal role of VDAC1 in mitochondrial integrity and cardiac function. Restoration of VDAC1-HK2 interaction prevented mPTP opening and cell death in microvascular ischemia-reperfusion injury. Liproxstatin-1 inhibited ferroptosis and promoted cardiomyocyte survival through reducing VDAC1 levels, but not VDAC2/3. Although VDAC1 is well known as a target for Parkin-mediated Lys 27 polyubiquitination and subsequent mitophagy, it can also be degraded by the ubiquitin-proteasome system. Our data confirmed that the protective effect of Parkin was obliterated by the proteasome inhibitor MG132, but not autophagy inhibitors. Recently, the proteolysis-targeting chimeras have demonstrated promise to selectively modulate protein content through the ubiquitin-proteasome system. Our findings favor the utility of targeting Parkin and VDAC1 for mitochondrial Ca2+ derangement and cardiac anomalies in obesity.

Study limitations

First, cisplatin was used to stimulate mitochondrial Ca2+ overload and mPTP opening, although it can also bind with DNA and evoke DNA damage culminating in mitochondrial apoptosis., Next, concurrent presence of diastolic and systolic dysfunction was observed in our HFD-induced obesity model. Given that diastolic dysfunction usually occurs prior to onset of systolic dysfunction, will deficiency in Parkin accelerate this process? Finally, Parkin is expressed in nearly all tissues, and a cardiac-specific knockout murine model would be more advantageous to directly address the role of Parkin in obesity cardiomyopathy in a more tissue-specific manner.

Conclusions

Findings from our present study provide evidence for the first time that Parkin deficiency accentuates HFD-induced cardiac anomalies but not aortic stiffness and BP change, likely through Parkin-VDAC1–dependent regulation of mitochondrial Ca2+ and mitochondrial integrity. Comprehensive evaluation of the role of Parkin in HFD-induced cardiac dysfunction rules out possible contribution from global metabolism on Parkin loss–accentuated cardiac anomalies. These findings favor the notion that Parkin, VDAC1, and mitochondria Ca2+ may serve as possible targets for drug development for heart dysfunction in patients with obesity. Further study is warranted to unveil the precise mechanism behind regulation of mitochondrial Ca2+ mobilization in a clinically relevant setting of obesity heart anomalies.

Funding Support and Author Disclosures

This work was supported by the National Natural Science Foundation of China (grants 82130011, 81900233, and 81770261). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.

COMPETENCY IN MEDICAL KNOWLEDGE: This study highlights that mitochondrial Ca2+ overload is one of the major candidates for obesity-induced mitochondria-mediated cell death, cardiac dysfunction, and remodeling. Activating Parkin-mediated degradation of VDAC1 in cardiomyocytes attenuates contractile dysfunction by limiting mitochondrial Ca2+ overload.

TRANSLATIONAL OUTLOOK: Future studies are warranted to determine the potential therapeutic benefit of Parkin activation and suppression of mitochondrial Ca2+ import, especially by promoting selective proteolysis of VDAC1, in cardiomyocytes for the prevention of cardiac remodeling and heart failure in the context of chronic obesity.